JP7163757B2 - Conversation processing system, information processing device and program - Google Patents

Description

The present invention relates to a dialogue processing system and related technology.

A database preliminarily registered in a server in which assumed questions (assumed questions) and answer information corresponding to the assumed questions are associated and defined in advance, and one assumed question corresponding to a question message from a user in an information processing device. There is a technology that acquires one answer information associated with from the server and notifies the user.

For example, in the technology described in Patent Document 1, when an information processing device receives a question message from a user, it transmits the question message to the server of the conversation processing system. In the database of the server, one answer message is registered in the answer information associated with the assumed question, and the server retrieves the answer information associated with the one assumed question corresponding to the question message from the database. It extracts and transmits to the information processing device. Then, the information processing device receives the response information (one response message) from the server and notifies the user of it.

JP 2016-126452 A

Here, for example, it is preferable to switch the answer message to be notified to the user depending on the employment form of the user (questioner who performed the questioning act) (whether the questioner is a regular employee or a temporary employee) . Also, it may be preferable to switch the answer message to be notified to the user according to the date (question date) when the question act is performed.

In this way, it may be preferable to switch the answer message to be notified to the user according to the information (also referred to as question act information) about the user's question act.

However, in the technique described in Patent Document 1, only one answer message is registered in the answer information associated with the assumed question corresponding to the question message. information) is notified to the user (questioner) as it is. Therefore, with the technique described in Patent Literature 1, it is not possible to switch the answer message notified to the user according to the question act information.

The operation of the server (such as the operation of extracting the answer information associated with the assumed question corresponding to the question message from the database and sending it back to the information processing device) depends on the provider of the conversation processing system (conversation processing service). is doing. Therefore, it is difficult for the user (administrator) of the conversation processing system to change (customize) the operation of the server.

Accordingly, it is an object of the present invention to provide a technology that enables more flexible responses to question messages from users.

In order to solve the above problems, the invention of claim 1 is a conversation processing system, which is a database defined by associating an information processing device, an assumed question that is an assumed question, and answer information corresponding to the assumed question. and a server having When received from the information processing device, one answer information associated with one assumed question corresponding to the question message is extracted from the database and transmitted to the information processing device, and the one answer information includes , a plurality of answer messages are defined for the one assumed question, and the information processing device includes receiving means for receiving the one answer information from the server; Acquisition means for acquiring action information, extraction means for extracting a single answer message corresponding to the question action information from among the plurality of answer messages in the one answer information, and extracting the single answer message and notification means for notifying the user , wherein each of the plurality of response messages is associated with at least one piece of attribute information indicating attributes different from each other in the one response information, and The extraction means extracts the attribute information corresponding to the question act information from the one answer information, and extracts the single answer message associated with the attribute information corresponding to the question act information from the plurality of answers. It is characterized by extracting from the message.

The invention of claim 2 is the conversation processing system according to the invention of claim 1 , wherein the information processing device stores a template of an inquiry destination notification message, which is a message for notifying an inquiry destination regarding the one assumed question. and the notifying means notifies the user of the inquiry destination notifying message based on the template in the information processing apparatus when the single reply message is not extracted by the extracting means.

The invention of claim 3 is the conversation processing system according to the invention of claim 2 , wherein the information processing device includes a plurality of inquiry destinations and a plurality of character strings for identifying each of the plurality of inquiry destinations, and are registered in association with each other, and the template of the inquiry destination notification message includes a predetermined changeable part indicating the inquiry destination, and the notification means stores the single reply message as the If not extracted by the extracting means, one specifying character string, which is one of the plurality of specifying character strings, is extracted from the one answer information, and the one specifying character string identifying one contact point corresponding to from among the plurality of contact points, changing the predetermined changeable portion of the template to the one contact point to generate the contact point notification message, and generating the contact point notification message; A notification message is notified to the user.

The invention of claim 4 is the conversation processing system according to any one of claims 1 to 3 , wherein the question act information includes classification information given to the questioner who performed the question act. It is characterized by

The invention of claim 5 is the conversation processing system according to the invention of claim 4 , wherein the classified information is information classified based on at least one of employment status and length of service at the company to which the questioner belongs. characterized by

The invention of claim 6 is the conversation processing system according to any one of claims 1 to 3 , wherein the question act information includes question date information indicating a date on which the question act was performed. and

The invention of claim 7 is the conversation processing system according to any one of claims 1 to 3 , wherein the question act information includes question time information indicating the time when the question act is performed. and

The invention of claim 8 is the conversation processing system according to any one of claims 1 to 3 , wherein the question act information includes question place information indicating a place where the question act is performed. and

According to the ninth aspect of the invention, a computer incorporated in an information processing apparatus is provided with a) a step of receiving a question message from a user, b) an assumed question that is an assumed question, and answer information corresponding to the assumed question. a step of sending the question message to a server having an associated defined database; a step of receiving from the server one answer information defining a plurality of answer messages; d) obtaining question act information that is information about question act by the user; e) corresponding to the question act information A program for executing the steps of: extracting a single reply message from among the plurality of reply messages in the one reply information; and f) notifying the user of the single reply message, wherein , each of the plurality of reply messages is associated with at least one piece of attribute information indicating a mutually different attribute of each reply message in the one reply information; attribute information is extracted from the one answer information, and the single answer message associated with the attribute information corresponding to the question act information is extracted from the plurality of answer messages. do.

The invention of claim 10 is the program according to the invention of claim 9 , wherein the information processing device stores a template of an inquiry destination notification message, which is a message for notifying an inquiry destination regarding the one assumed question, In the step f), if the single reply message is not extracted in the step e), the user is notified of the contact notification message based on the template in the information processing device. do.

The invention of claim 11 is the program according to the invention of claim 10 , wherein in the information processing device, a plurality of inquiry destinations and a plurality of identifying character strings for identifying each of the plurality of inquiry destinations are respectively The template of the inquiry destination notification message is associated and registered, and includes a predetermined changeable part indicating the inquiry destination, and in step f), the single reply message is the If not extracted in step e), one specifying character string, which is one of the plurality of specifying character strings, is extracted from the one answer information, and the one specifying character string is extracted from the one answer information. one contact point corresponding to the character string is specified from among the plurality of contact points, the modifiable part of the template is changed to the one contact point, the contact point notification message is generated, and the contact point is generated. A notification message is notified to the user.

The invention of claim 12 is the program according to any one of claims 9 to 11 , wherein the question act information includes classification information given to the questioner who performed the question act. Characterized by

The invention of claim 13 is the program according to the invention of claim 12 , wherein the classification information is information classified based on at least one of employment status and length of service at the company to which the questioner belongs. and

A fourteenth aspect of the invention is the program according to any one of the ninth to eleventh aspects of the invention, wherein the question act information includes question date information indicating a date when the question act is performed. .

According to a fifteenth aspect of the invention, in the program according to any one of the ninth to eleventh aspects of the invention, the question act information includes question time information indicating a time when the question act is performed. .

A sixteenth aspect of the invention is the program according to any one of the ninth to eleventh aspects of the invention, wherein the question act information includes question place information indicating a place where the question act is performed. .

A seventeenth aspect of the present invention is an information processing apparatus comprising: a receiving means for receiving a question message from a user; a transmission means for transmitting the question message to a server having a server; Receiving means for receiving only one answer information from the server, acquisition means for acquiring question act information that is information on question act by the user, and a single answer message corresponding to the question act information to the one extracting means for extracting from among the plurality of reply messages in reply information; and notification means for notifying the user of the single reply message , wherein each of the plurality of reply messages is the one The answer information is associated with at least one attribute information indicating mutually different attributes of each answer message, and the extracting means extracts the attribute information corresponding to the question act information from the one answer information, The single answer message associated with the attribute information corresponding to the question act information is extracted from the plurality of answer messages.

According to the invention described in claims 1 to 17 , when one answer information associated with one assumed question corresponding to the question message from the user is received from the server, the question about the question action by the user is received. A single reply message corresponding to the action information is extracted from among the plurality of reply messages in the one reply information. The single reply message is then notified to the user. Therefore, it is possible to flexibly reply to the question message from the user.

In particular, according to the second and tenth aspects of the invention, the information processing apparatus stores a template of the inquiry destination notification message, and when a single response message is not extracted, the template in the information processing apparatus The user is notified of the inquiry destination notification message based on. Therefore, it is possible to respond more flexibly to the question message from the user.

1 is a schematic diagram showing the configuration of a conversation processing system; FIG. It is a figure which shows the functional block of an information processing apparatus. It is a figure which shows the schematic operation|movement of a conversation processing system. It is a figure which shows a chat screen (before inputting a question message). It is a figure which shows a chat screen (after inputting a question message). It is a flowchart which shows operation|movement (answer extraction operation|movement etc.) of an information processing apparatus. It is a timing chart which shows an example of operation|movement of a conversation processing system. It is a figure which shows an answer management table. It is a figure which shows specific answer information. It is a figure which shows a classification information management table. It is a figure which shows a chat screen (after an answer notification). It is a figure which shows the chat screen (after inputting a question message) which concerns on 2nd Embodiment. It is a figure which shows the reply management table based on 2nd Embodiment. It is a figure which shows the chat screen (after notification of an answer message) which concerns on 2nd Embodiment. 14 is a flow chart showing operations (answer extraction operations, etc.) of the information processing apparatus according to the third embodiment. It is a figure which shows the reply management table based on 3rd Embodiment. It is a figure which shows an inquiry destination management table. It is a figure which shows the chat screen (after notification of an answer message) which concerns on 3rd Embodiment. It is a figure which shows the reply management table based on a modification. It is a figure which shows the classification information which concerns on a modification.

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings.

<1. First Embodiment>
<1-1. System configuration>
FIG. 1 is a schematic diagram showing the configuration of a conversation processing system 1 according to the present invention. As shown in FIG. 1 , the conversation processing system 1 includes a computer 50 (information processing device used by users) and a server 90 .

Elements 50 and 90 in the system 1 are communicably connected to each other via a network 108 . The network 108 is configured by a LAN (Local Area Network), the Internet, and the like. Also, the mode of connection to the network 108 may be wired connection or wireless connection.

The server 90 is an external device (external server) provided separately from the computer 50 . The server 90 responds to the computer 50 (the user) with answer information A100 (see FIG. 3, etc.) associated with an assumed question (question expected to be asked by the user) corresponding to the question message Q10 from the user. It is a server that provides a conversation processing service for The server 90 has an automatic dialogue program or the like that establishes a pseudo-conversation (chat) with the user, and uses the automatic dialogue program or the like to return the reply information A100 to the computer 50 . The server 90 is also called a chatbot server or the like. The conversation processing system 1 is also called a chatbot system or the like.

<1-2. Configuration of Computer 50>
FIG. 2 is a functional block diagram showing a schematic configuration of the computer 50. As shown in FIG.

The computer 50 is an information input/output terminal device (information processing device) capable of network communication with other devices (chatbot server 90, etc.). Here, a smartphone is exemplified as the computer 50 . However, the computer 50 is not limited to this, and may be a tablet terminal or a so-called personal computer.

The computer 50, as shown in the functional block diagram of FIG. , to implement various functions.

The communication unit 54 can perform network communication via a network. Various protocols such as TCP/IP (Transmission Control Protocol/Internet Protocol) are used in this network communication. By using the network communication, the computer 50 can exchange various data with a desired partner (chatbot server 90, etc.). The communication unit 54 has a transmission unit 54a that transmits various data and a reception unit 54b that receives various data.

The storage unit 55 is configured by a storage device such as a hard disk drive (HDD) and/or a semiconductor memory.

The operation unit 56 includes an operation input unit 56a that receives operation input to the computer 50, and a display unit 56b that displays and outputs various information (chat screen 200 (FIG. 4), etc.).

A controller (control unit) 59 in FIG. 2 is a control device that is built in the computer 50 and controls the computer 50 in an integrated manner. The controller 59 is configured as a computer system including a CPU and various semiconductor memories (RAM and ROM). The controller 59 implements various processing units by executing a predetermined program stored in a storage unit (semiconductor memory or the like) in the CPU. The program (specifically, the program module group) may be recorded in a portable recording medium such as a USB memory, read out from the recording medium, and installed in the computer 50 . Alternatively, the program may be downloaded via the network 108 or the like and installed on the computer 50 .

The computer 50 is installed with an application software program (also referred to as conversation processing application) capable of answering a question message Q10 from the user in an interactive format (chat format). The conversation processing application realizes the reply operation by cooperating with the chatbot server 90 .

The controller 59 realizes various processing units including the communication control unit 41, the input control unit 42, the display control unit 43, the extraction unit 44, the acquisition unit 45, and the determination unit 46 by executing the programs and the like.

The communication control unit 41 is a processing unit that cooperates with the communication unit 54 and the like to control communication operations with the chatbot server 90 and the like. For example, the communication control unit 41 cooperates with the communication unit 54 (more specifically, the transmission unit 54 a ) and the like to transmit the question message Q<b>10 from the user to the chatbot server 90 . In addition, the communication control unit 41 cooperates with the communication unit 54 (more specifically, the receiving unit 54b) and the like to receive the answer information A100 associated with the assumed question Q50 corresponding to the question message Q10 from the chatbot server 90. do.

The input control unit 42 is a control unit that controls an operation (operation input operation) of receiving an operation input by the user in cooperation with the operation input unit 56a. For example, the input control unit 42 cooperates with the operation input unit 56a to receive a question message Q10 (input operation of the question message Q10) from the user on the chat screen 200 (FIG. 4, etc.).

The display control unit 43 is a processing unit that controls the display operation of various information (chat screen 200 (FIG. 4), etc.) on the display unit 56b (FIG. 1). For example, the display control unit 43 uses the chat screen 200 to notify (present) the user of an answer message to the question message Q10 from the user.

The extraction unit 44 extracts a single answer message corresponding to the question act information (described below) from among the plurality of answer messages 400 (see FIG. 8) in the answer information A100 received from the chatbot server 90. It is a processing unit that executes processing (hereinafter also referred to as reply message extraction processing).

The acquisition unit 45 is a processing unit that controls an acquisition operation for acquiring information (question act information) about the question act by the user.

The determination unit 46 is a processing unit that executes various determination processes.

Here, the various operations described above are executed mainly by the CPU of the controller 59 executing the software program, but the present invention is not limited to this. The various operations described above may be executed using dedicated hardware or the like provided internally or externally. For example, all or part of the communication control unit 41, the input control unit 42, the display control unit 43, the extraction unit 44, the acquisition unit 45, the determination unit 46 (FIG. 2), and the like use one or more dedicated hardware. may be implemented.

<1-3. Reply Management Table 300>
Before explaining the operation of the conversation processing system 1, the reply management table 300 (see FIG. 8) will be explained.

The chatbot server 90 stores a response management table 300 (301). FIG. 8 is a diagram showing the reply management table 301 (300).

In the answer management table 300, an assumed question (assumed question) Q50 and answer information A100 corresponding to the assumed question Q50 are defined (registered) in association with each other.

Here, a plurality of assumed questions Q51 to Q53 and a plurality of pieces of answer information A101 to A103 corresponding to the plurality of assumed questions Q51 to Q53 are respectively associated and defined in the answer management table 301 (see FIG. 8). .

Specifically, in the answer management table 301, an assumed question Q51 (question regarding a request for commuting by car) and answer information A101 corresponding to the assumed question Q51 are defined in association with each other. Further, in the answer management table 301, an assumed question Q52 (question related to a translation request) and answer information A102 corresponding to the assumed question Q52 are defined in association with each other. Furthermore, in the answer management table 301, an assumed question Q53 (question regarding browsing of the bus timetable) is defined in association with answer information A103 corresponding to the assumed question Q53. Each of the answer information A100 (A101 to A103) is composed of text data.

In addition, in this conversation processing system 1, a plurality of users are classified in advance into a plurality of employee types (a plurality of employee type groups) based on the employment status and length of service in the company to which each user of the conversation processing system 1 belongs. It is Specifically, there are three types of users: regular employees who have been with the company for two years or more (regular employees who have been working for one year or more), regular employees who have been with the company for one year or more (regular employees who have been working for less than one year), and temporary employees. They are classified into employee types (three employee type groups "G1", "G2", and "G3"). More specifically, regular employees who have been with the company for two years or more are classified into the employee type group G1, regular employees who have been with the company for the first year are classified into the employee type group G2, and temporary workers are classified into the employee type group G3.

In the reply management table 300 (301), respective reply information A101 to A103 are defined in advance based on the three employee types (employee types). Specifically, in each of the reply information A101 to A103, reply messages 400 corresponding to the three employee types (three employee type groups "G1", "G2", and "G3") are defined in advance. These prescribed contents may be registered in the reply management table 300 by the manager of the conversation processing system 1 or the like.

In the answer information A101 (answer information corresponding to the assumed question Q51), three answer messages 411 to 413 (see FIG. 9) corresponding to the three employee types are defined as answer messages to the assumed question Q51. there is FIG. 9 is a diagram showing reply information A101.

The reply message 411 is for regular employees who have been with the company for two years or more (for users who belong to the employee type group G1). Here, the reply message 411 is composed of a character string "Parking lot usage application is required."

The reply message 412 is a message for full-time employees in their first year of employment (for users belonging to the employee type group G2). Here, the reply message 412 is composed of a character string of "I am not qualified to use the parking lot. Please check the work rules."

The reply message 413 is a message for temporary staff (for users belonging to the employee type group G3). Here, the reply message 413 is composed of a character string of "Parking is not available."

Thus, in this conversation processing system 1, not only one reply message is defined in one reply information A100, but a plurality of reply messages 411 to 413 are defined in one reply information A100 (A101). .

Further, in the reply information A101 (FIG. 9), the plurality of reply messages 411 to 413 and the plurality of attribute information 500 (501 to 503) are defined in association with each other. Attribute information (attribute data) 500 (501 to 503) is information (data) indicating attributes of each of the reply messages 411 to 413 (more specifically, classification attributes for classifying the reply message 400).

The attribute information 501 is information indicating that the reply message associated with the attribute information 501 is for regular employees who have been with the company for two years or more (for users belonging to the employee type group G1). Here, the same character string “G1” as the group name of the employee type group G1 is used as the attribute information 501 .

The attribute information 502 is information indicating that the reply message associated with the attribute information 502 is intended for full-time employees in their first year of employment (for users belonging to the employee type group G2). Here, the same character string “G2” as the group name of the employee type group G2 is used as the attribute information 502 .

The attribute information 503 is information indicating that the reply message associated with the attribute information 503 is for temporary staff (for users belonging to the employee type group G3). Here, the same character string “G3” as the group name of the employee type group G3 is used as the attribute information 503 .

In the answer management table 301, each attribute information 501-503 and each answer message 411-413 are associated by defining each attribute information 501-503 and each answer message 411-413 as a series of character strings. . Since each series of character strings includes each attribute information 500 in addition to each reply message (answer character string) 400, it is also called an expanded reply character string M100 (see FIG. 9). That is, each extended character string for reply M100 is composed of each attribute information 500 and each character string for reply (here, reply message 400).

Specifically, the extended response character string M111 (M100) is a series of character strings including the attribute information 501 and the response message 411, and the extended response character string M112 (M100) includes the attribute information 502 and the response message. message 412 is a series of strings. Further, the extended response character string M113 (M100) is a series of character strings including the attribute information 503 and the response message 413. FIG.

In addition, as shown in FIG. 9, two types of distinguishing information (also called delimiters) D10 and D20 are also defined in each extended answer character string M100.

The distinguishing information D10 is information for distinguishing a plurality of extended answer character strings M100 (M111 to M113) from each other. A predetermined type of symbol (code) is used as the discrimination information D10. Here, square brackets are used as the distinguishing information D10.

The distinguishing information D20 is information for distinguishing the reply message 400 and the attribute information 500 in each expanded character string for reply M100 (M111 to M113). As the distinguishing information D20, a type of symbol (code) different from that of the distinguishing information D10 is used. Here, a closing square bracket is used as the distinguishing information D20.

Thus, in the reply information A101 (FIG. 8), three reply expanded character strings M111-M113 containing three reply messages 411-413 and three pieces of attribute information 501-503 are defined.

Also, in the answer information A102 (answer information corresponding to the assumed question Q52) (FIG. 8), two expanded character strings for answer M121 and M122 (M100) are defined.

In the reply extended character string M121, two pieces of attribute information 501 and 502 are associated with one reply message 421 and defined. Specifically, in the reply extended character string M121, the reply message 421 and attribute information 501 are defined in association, and the reply message 421 and attribute information 502 are defined in association. The reply message 421 is for regular employees who have been with the company for two years or more and for regular employees who have been with the company for the first year (that is, for regular employees). In other words, the reply message 421 is a message for users who belong to either the employee group G1 or G2. Here, the reply message 421 is composed of a character string of "here is the request method. <https://*****/>".

In the reply extended character string M122, the reply message 422 and the attribute information 503 are defined in association with each other. The reply message 422 is a message for temporary workers. In other words, the reply message 422 is a message for users belonging to the employee type group G3. Here, the reply message 422 is composed of a character string "Please consult with the full-time employee in charge."

In answer information A103 (answer information corresponding to assumed question Q53), attribute information 500 is not specified, and answer message 431 (only one answer message 431) is specified as an answer message to assumed question Q53. The reply message 431 is a message for all employees (both regular employees and temporary employees). Here, the reply message 431 is composed of a character string of "This is it. <https://**********/>".

An answer management table 301 (300) defining the answer information A100 as described above is registered in the chatbot server 90 in advance.

Then, as will be described later, the chatbot server 90 extracts the answer information A100 associated with one assumed question Q50 corresponding to the question message Q10 from the user from the answer management table 301 (FIG. 8). Send to computer 50 .

<1-4. Operation>
Next, the operation of the conversation processing system 1 will be explained. FIG. 3 is a diagram showing a schematic operation of the conversation processing system 1. As shown in FIG.

The computer 50 receives the question message Q10 from the user and transmits the question message Q10 to the chatbot server 90 . As described above, the chatbot server 90 is registered in advance in association with the answer information A100 corresponding to the assumed question Q50 (see FIG. 8). Also, in the answer information A100, a plurality of answer messages 400 (411 to 413) are defined in advance for the assumed question Q50.

When the question message Q10 is received from the computer 50, the chatbot server 90 transmits (replies) to the computer 50 the answer information A100 associated with the assumed question Q50 corresponding to the question message Q10. The computer 50 receives the answer information A100 and acquires information (question action information) on the question action by the user. Then, the computer 50 extracts a single reply message (412) corresponding to the question act information from among the multiple reply messages 411 to 413 in the reply information A100 and notifies the user of it.

Details of such operations are described below.

First, the computer 50 activates the conversation processing application in response to an activation operation by the user (here, user U1).

When the activation of the conversation processing application is completed, the computer 50 displays a user information input screen (not shown) on the display unit 54b (FIG. 1), and user information (user identification information) (user ID etc.) from the user (U1). The user information is used in the operation of obtaining question act information (step S15), as will be described later. When the user information input operation is accepted, the computer 50 displays a chat screen 200 (201) on the display unit 54b (see FIG. 4).

Also, the computer 50 starts the operation of FIG. 6 in response to the completion of activation of the conversation processing application. FIG. 6 is a flow chart showing the operation of the computer 50 (answer extraction operation, etc.). 7 is a timing chart showing an example of the operation of the conversation processing system 1. As shown in FIG.

First, the computer 50 receives a question message Q10 from the user U1 on the chat screen 200 (201) (step S11). Here, user U1 inputs a question message Q11 ("I want to commute by car") on chat screen 201 (see FIG. 5), and computer 50 receives question message Q11 from user U1.

The computer 50 then transmits the question message Q11 input by the user U1 to the chatbot server 90 (step S12) (see also FIG. 3).

When the chatbot server 90 receives the question message Q11 from the computer 50 (step S51 (see FIG. 7)), it analyzes the question message Q11 (contents of the question message Q11) using natural language analysis processing or the like (step S52).

Then, the chatbot server 90 extracts the answer information A100 associated with the assumed question Q50 corresponding to the question message Q11 from the answer management table 301 (300) (see FIG. 8) (step S53).

Specifically, the chatbot server 90 uses natural language analysis processing or the like to convert one assumed question Q50 corresponding to the question message Q11 from the plurality of assumed questions Q51 to Q53 (FIG. 8) in the answer management table 301. Extract from inside. Here, a hypothetical question Q51 regarding a request for commuting by car is extracted. Then, the chatbot server 90 extracts the answer information A100 (here, the answer information A101) associated with the assumed question Q51 from the answer management table 301. FIG.

After that, the chatbot server 90 transmits the answer information A101 to the computer 50 (step S54). Specifically, instead of extracting one of the plurality of reply messages 411 to 413 from the reply information A101 and transmitting it to the computer 50 in the chatbot server 90, the entire reply information A101 (a plurality of Reply information A101) including reply messages 411-413 is sent to the computer 50. FIG.

When the computer 50 receives the answer information A101 from the chatbot server 90 (step S13 (FIG. 6)), based on the distinguishing information D10 (see FIG. 9), a plurality of answer messages 400 (a plurality of It is determined whether or not the extended character string for answer M100) exists (step S14).

For example, if multiple pieces of distinguishing information D10 exist in the answer information A101, the computer 50 determines in step S14 that multiple answer messages 400 exist in the answer information A101. Here, it is determined in step S14 that there are three pieces of distinguishing information D10 in the answer information A101 (see FIG. 8), and that there are three answer messages 400 in the answer information A101. Then, the process proceeds from step S14 to step S15.

If the distinguishing information D10 does not exist in the reply information A100, the fact that a plurality of reply messages 400 do not exist in the reply information A101 (the fact that only a single reply message exists in the reply information A101) is notified in step S14. , and the process proceeds from step S14 to step S17 (described later). For example, when the question message Q10 requesting to see the bus timetable is received, the answer information A103 (answer information without the distinction information D10) (see FIG. 8) associated with the assumed question Q53 is sent to the chatbot server 90. to the computer 50 (step S54). In this case, it is determined in step S14 that multiple reply messages 400 do not exist in the reply information A103, and the process proceeds from step S14 to step S17.

In step S15, the computer 50 acquires question act information regarding the question act by the user (more specifically, the subject of the question act (questioner)). Here, the classification information given in advance (before the question message Q11 is accepted) to the questioner (questioner) who performed the questioning action is acquired as the questioning action information. Specifically, the computer 50 classifies the employee type information (employee type group) of the user U1 (questioner) based on the user information (user ID) input on the user information input screen (not shown). Get it as information.

More specifically, the conversation processing system 1 also has a classification information management server (not shown). The classification information management server stores a classification information management table 900 (901) (see FIG. 10). FIG. 10 is a diagram showing the classification information management table 901. As shown in FIG.

In the classification information management table 901, the user information (user ID) of each user of the conversation processing system 1 and the classification information of each user (specifically, employee type groups G1 to G3) are associated and registered. there is As described above, regular employees who have been with the company for two years or more are classified into the employee type group G1 based on the type of employment and length of service at the company to which the user using the conversation processing system 1 belongs. Regular employees are classified into the employee type group G2, and temporary employees are classified into the employee type group G3.

The computer 50 transmits the user information of user U1 (“userU1”) to the classification information management server. When receiving the user information of user U1 from the computer 50, the classification information management server extracts classification information (employee type group (here, G2)) corresponding to the user information from the classification information management table 901. FIG. In other words, an employee type group given to user U1 based on the employment type and length of service of user U1 is extracted. Then, the classification information management server transmits the extracted classification information (information indicating the group to which the user U1 (questioner) belongs among the plurality of employee type groups G1 to G3) to the computer 50, and the computer 50 transmits the classification Information is received and acquired from the classification information management server (step S15).

Then, the process proceeds from step S15 to step S16.

In step S<b>16 , the computer 50 executes answer message extraction processing based on the question act information (specifically, the employee type group G<b>2 of the user U<b>1 ) and the answer information A<b>101 from the chatbot server 90 . The answer message extraction process is a process of extracting a single answer message (an answer message to be answered to the questioner) from among the plurality of answer messages 411 to 413 in the answer information A101.

Specifically, the computer 50 extracts the answer message 400 corresponding to the question act information (employee type group G2 here) from the answer messages 411 to 413 in the answer information A101.

More specifically, the computer 50 first extracts the attribute information 500 corresponding to the question act information from the answer information A101. Specifically, the computer 50 extracts a plurality of expanded answer character strings M111 to M113 from the answer information A101 using the distinguishing information D10 (see FIG. 9), and extracts the distinguishing information D20 (see FIG. 9). is used to extract a plurality of pieces of attribute information 500 (501 to 503) from the extracted extended answer character strings M111 to M113. Then, the computer 50 extracts attribute information 500 corresponding to the question act information (employee type group G2) from among the plurality of extracted attribute information 501-503. Here, the attribute information 502 (see FIG. 9) having the same character string "G2" as the employee type group G2 is extracted.

Then, the computer 50 generates a plurality of reply messages 412 (that is, reply messages 412 for full-time employees (for users belonging to the employee type group G2) in the first year of employment) associated with the attribute information 502 in the reply information A101. Extract from messages 411-413. Specifically, the computer 50 extracts the reply message 412 associated with the attribute information 502 in the reply extended character string M112 containing the attribute information 502. FIG.

After that, the process proceeds from step S16 to step S17, and the computer 50 notifies the user U1 (questioner) of the answer message 412 extracted by the answer message extraction process (step S16). Specifically, the computer 50 displays the answer message 412 (“You are not qualified to use the parking lot. Please check the work rules.”) on the chat screen 201 (200) (see FIG. 11).

As described above, in the first embodiment, a plurality of reply messages 411 to 413 are defined in advance in the reply information A101 in the reply management table 301 of the chatbot server 90 (see FIG. 8). When the computer 50 receives the answer information A101 associated with the assumed question Q50 corresponding to the question message Q10 from the user from the chatbot server 90 (step S13), it sends a single answer message 412 corresponding to the question act information. , are extracted from the plurality of reply messages 411 to 413 in the reply information A101 (step S16). Then, the single reply message 412 is notified to the user (step S17) (see also FIG. 11). Therefore, since the answer message 400 notified to the user is switched according to the question act information (here, the classification information given to the user), it is possible to flexibly answer the question message Q10 from the user. is possible.

In particular, in the above-described first embodiment, the chatbot server 90 provides one answer associated with the assumed question Q50 corresponding to the question message Q10 from the user, regardless of the defined contents of the answer information A100 in the answer management table 300. The information A100 is sent to the computer 50, and the computer 50 receives the one reply information A100. However, the one reply information A100 (for example, A101) does not define only one reply message, but predefines a plurality of reply messages 400 (411 to 413) (see FIG. 8). Then, the computer 50 extracts a single reply message (412) corresponding to the question act information from among the multiple reply messages 411 to 413 in the reply information A101 and notifies the user of it. Therefore, without changing the operation of the chatbot server 90 itself (the operation of extracting one answer information A100 corresponding to the assumed question Q50 of the question message Q10 from the answer management table 300 and replying to the computer 50), the user It is possible to implement a flexible answer to the question message Q10 from .

In the first embodiment, the computer 50 receives the answer information A100 from the chatbot server 90 (step S13) and then acquires the question act information (employee type group) (step S15). is not limited to For example, the question act information may be acquired before answer information A100 is received from chatbot server 90 (for example, immediately after user information is input on the user information input screen (immediately before step S11)).

<2. Second Embodiment>
The second embodiment is a modification of the first embodiment. Below, it demonstrates centering around difference with 1st Embodiment.

In the above-described first embodiment, the classification information (information about the subject of the questioning act) given to the questioner who performed the questioning act is acquired as the questioning act information.

On the other hand, in the second embodiment, question date information (information about the timing of the question act) indicating the date of the question act is acquired as the question act information.

The operation of FIG. 6 is executed in the computer 50 also in this second embodiment. Here, it is assumed that user U1 inputs a question message Q14 ("When will the next power outage occur?") on chat screen 200 (202) (see FIG. 12). Also, here, it is assumed that the question date by the user U1 is July 31, 2018.

When the computer 50 receives the question message Q14 (Q10) from the user U1 (step S11), the computer 50 transmits the question message Q14 to the chatbot server 90 (step S12).

When the chatbot server 90 receives the question message Q14 from the computer 50 (step S51 (FIG. 7)), the chatbot server 90 extracts the answer information A100 associated with the assumed question Q50 corresponding to the question message Q14 from the answer management table 300. Extract (step S53).

FIG. 13 is a diagram showing the reply management table 300 (302) according to the second embodiment.

Here, two assumed questions Q54 and Q55 and two pieces of answer information A104 and A105 corresponding to the two assumed questions Q54 and Q55 are associated and defined in the answer management table 302 . Specifically, in the answer management table 302, an assumed question Q54 (question regarding power outage day) and answer information A104 (A100) corresponding to the assumed question Q54 are defined in association with each other. In addition, in the answer management table 302, the assumed question Q55 (question regarding the evacuation drill date) and the answer information A105 (A100) corresponding to the assumed question Q55 are defined in association with each other.

In the response information A104 (A100), two extended response character strings M141 and M142 (M100) are included based on the occurrence of power outages on August 1, 2018 and August 17, 2018, respectively. stipulated.

In the reply extended character string M141 (M100), the reply message 441 (400) and the attribute information 511 (500) are defined in association with each other. The reply message 441 is a message to the effect that a power outage will occur on August 1st. Here, the reply message 441 is composed of a character string "There will be a power outage at the Tokyo site on August 1.". Also, date information “2018.8.1” indicating the power outage date indicated in the reply message 441 is defined as the attribute information 511 .

In the reply extended character string M142 (M100), the reply message 442 (400) and the attribute information 512 (500) are defined in association with each other. The reply message 442 is a message to the effect that a power outage will occur on August 17th. Here, the response message 442 is composed of a character string "There will be a power outage at the Osaka site on August 17." Also, date information “2018.8.17”, which is the same as the power outage date indicated in the reply message 442 , is specified as the attribute information 512 .

In response information A105 (A100), based on the evacuation drills being held on October 1, 2018 and December 15, 2018, respectively, two extended response character strings M151 and M152 (M100 ) are stipulated.

In the reply extended character string M151 (M100), the reply message 451 (400) and the attribute information 521 (500) are defined in association with each other. The reply message 451 is a message to the effect that an evacuation drill will be held on October 1, 2018. Here, the response message 451 is composed of a character string "There will be an evacuation drill for an earthquake on October 1." Also, date information “2018.10.1” indicating the evacuation drill date indicated in the reply message 451 is defined as the attribute information 521 .

In the reply extended character string M152 (M100), the reply message 452 (400) and the attribute information 522 (500) are defined in association with each other. The reply message 452 is a message to the effect that an evacuation drill will be held on December 15, 2018. Here, the response message 452 is composed of a character string of "There will be an evacuation drill in case of fire on December 15th." Also, date information “2018.12.15” indicating the evacuation drill date indicated in the reply message 452 is defined as the attribute information 522 .

The chatbot server 90 extracts the answer information A100 associated with the assumed question Q50 corresponding to the question message Q14 from the answer management table 302 (FIG. 13) (step S53) and transmits it to the computer 50 (step S54). ).

Specifically, first, when the chatbot server 90 receives the question message Q14 from the computer 50 (step S51 (FIG. 7)), it analyzes the question message Q14 using natural language analysis processing or the like (step S52 ).

Then, the chatbot server 90 extracts the assumed question Q54 (question Q54 regarding the power outage date) corresponding to the question message Q14 from among the plurality of assumed questions Q54 and Q55, and also extracts the answer information associated with the assumed question Q54. A104 is extracted from the answer management table 302 (step S53). After that, the chatbot server 90 transmits the answer information A104 to the computer 50 (step S54).

When the computer 50 receives the answer information A104 from the chatbot server 90 (step S13 (FIG. 6)), based on the presence of a plurality of distinguishing information D10 (FIG. 13) in the answer information A104, the In step S14, it is determined that a plurality of reply messages 400 exist in the reply information A104.

Then, the process proceeds from step S14 to step S15, and the computer 50 acquires question act information (step S15). Specifically, the computer 50 has a function of acquiring the input date (here, July 31, 2018) of the question message Q14, and the date (July 2018) on which the questioning action was performed by the user U1. 31st of the month) is obtained as question act information.

Thereafter, the computer 50 executes answer message extraction processing based on the question act information (more specifically, question date information) and the answer information A104 from the chatbot server 90 (step S16).

Specifically, the computer 50 uses the distinguishing information D10 (see FIG. 13) to extract the two expanded answer character strings M141 and M142 from the answer information A101, and extracts the distinguishing information D20 (see FIG. 13). ) is used to extract two pieces of attribute information 511 and 512 from the extracted extended character strings for answer M141 and M142. Then, the computer 50 extracts the attribute information 500 corresponding to the question date information (July 31, 2018) from the extracted two pieces of attribute information 511 and 512 . Specifically, the computer 50 selects the attribute information 511 (“2018.8.1”) indicating the most recent date after the question date information (July 31, 2018) from among the plurality of attribute information 511 and 512. Extract. Then, the computer 50 selects the answer message 441 associated with the attribute information 511 in the answer information A104 (more specifically, the extended answer character string M141 including the attribute information 511). Extract from

After that, the process proceeds from step S16 to step S17, and the computer 50 notifies the extracted answer message 441 to the user U1 (questioner). Specifically, the computer 50 displays the answer message 441 (“There will be a power outage at the Tokyo site on August 1.”) on the chat screen 202 (see FIG. 14).

As described above, in the second embodiment, the question date information indicating the date when the question act is performed is acquired as the question act information.

In addition, it is not limited to this, and the question location information indicating the place where the question act is performed may be acquired as the question act information.

Here, it is assumed that a reply message to the effect that there is a power outage at the Tokyo site and a reply message to the effect that there is a power outage at the Osaka site are specified in the reply information A100 (see also FIG. 13). In this case, as the attribute information 500, the blackout location (Tokyo site, Osaka site) indicated in each reply message 400 is defined. Specifically, there are two M100 extended response character strings (“[Tokyo site] There will be a power outage at the Tokyo site on August 1st.” and “[Osaka site] There will be a power outage at the Osaka site on August 17th.”). ”) is defined in the answer information A100.

For example, when the user inputs a question message Q14 (see FIG. 12) at the Tokyo site, the computer 50 acquires the location information of the user (location information at the time of receiving the question message Q10), and acquires the location information (the Based on the location information at the time of receiving the question message Q10), the question location information (here, the Tokyo site) is acquired as the question act information. Then, the computer 50 extracts an answer message (“There will be a power outage at the Tokyo site on August 1”) corresponding to the question location information (Tokyo site) and notifies the user of it.

In this way, question location information may be acquired as question act information.

Alternatively, question time information indicating the time at which the question act was performed may be acquired as the question act information.

Here, the response information A100 contains a response message to the effect that the conference will start at 10:00 in a certain conference room and a response message to the effect that the conference will start at 15:00 in the certain conference room. Assuming it is specified. Note that the attribute information 500 in this case defines the meeting start time indicated in each reply message. Specifically, two extended response character strings M100 (“[10:00] There is a meeting starting at 10:00”, “[15:00] There is a meeting starting at 15:00 Yes.") is defined in the answer information A100.

For example, when a user who wants to use a certain conference room enters a question message Q10 at 13:00 asking the start time of the next conference in the certain conference room, the computer 50 outputs the question time information (here Then 13:00) is acquired as question act information. Then, the computer 50 displays the answer message (“There is a meeting starting at 15:00”) associated with the attribute information 500 indicating the latest time after the question time information (13:00). Extract and notify the user.

In this way, question time information may be acquired as question act information.

<3. Third Embodiment>
The third embodiment is a modification of the first embodiment. Below, it demonstrates centering around difference with 1st Embodiment.

In the first embodiment, a single answer message is extracted from a plurality of answer messages 400 (a plurality of answer messages 400 in the answer information A100 defined in the chatbot server 90) and sent to the user (questioner) have been notified.

Here, a single reply message may not be extracted from among the multiple reply messages 400 . For example, if the questioner does not belong to any of the multiple employee type groups G1 to G3, there is no response message 400 corresponding to the question act information (employee type group), and a single response message 400 out of the multiple response messages 400 response message is not extracted.

In this third embodiment, in such a case, the computer 50 (conversation processing application) generates an inquiry destination notification message 700 (described below) and notifies the user of the exceptional processing (response information in the chatbot server 90). It is executed as a process (exception process) to the process of notifying the user of the answer message 400 defined in A100 (principle process).

The inquiry destination notification message 700 is a message for notifying the user (questioner) of the inquiry destination regarding the assumed question Q50 corresponding to the question message Q10 from the user. Here, the name of the person in charge of the business is notified as the inquiry destination, and the inquiry destination notification message 700 is also called a person-in-charge notification message or the like. However, it is not limited to this, and the name of the department in charge or the telephone number of the person in charge may be notified as the inquiry destination.

FIG. 16 is a diagram showing the reply management table 300 (303) according to the third embodiment.

In the answer management table 303, as in the first embodiment (answer management table 301 (FIG. 8)), three assumed questions Q51 to Q53 and three pieces of answer information A100 corresponding to the three assumed questions Q51 to Q53 and are defined in association with each other.

However, in the answer management table 303, the answer information A111 (A100) further defines an expanded answer character string M114 in addition to the three expanded answer character strings M111 to M113.

The expanded answer character string M114 (M100) is also composed of attribute information 500 and an answer character string, like the other expanded answer character strings M111 to M113.

However, in the extended response character string M114, an inquiry destination identification character string 601 (“parking management”) is defined as the response character string instead of the response message 400. FIG. The inquiry contact identification character string 601 (600) is a character string (answer character string) for identifying (extracting) the inquiry contact regarding the assumed question Q51 in the inquiry contact management table 800 (described below) (see FIG. 17). be. Here, as the inquiry destination identification character string 601, the business name (“parking lot management”) of the inquiry destination regarding the assumed question Q51 is defined.

Further, in the extended character string for answer M114, the character string "Info" is defined as the attribute information 550 (500) in association with the character string for answer (601). The character string "Info" indicates that the following character string (here, "parking management") is an exception processing character string (a character string used for exception processing (notification processing of the contact notification message 700)). showing. In other words, the character string "Info" is attribute information indicating that the subsequent answer character string (601) is information for exception handling ("exception handling information").

FIG. 17 is a diagram showing an inquiry destination management table 800 and the like. In the inquiry destination management table 800, a person in charge (inquiry destination) for each business content (business type) is defined. The inquiry destination management table 800 is stored in the computer 50 .

In the inquiry destination management table 800, a plurality of inquiry destinations (here, a plurality of persons in charge) and a plurality of inquiry destination identification character strings for identifying each of the plurality of inquiry destinations are registered in association with each other. ing. The inquiry destination management table 800 is also called an inquiry destination identification table.

In the inquiry destination management table 800, a plurality of persons in charge are registered as a plurality of inquiry destinations, and a plurality of tasks (task types) are registered as a plurality of inquiry destination specifying character strings. Specifically, each business type and the person in charge of each business are registered in association with each other. That is, a person in charge of each task is registered (for each task).

For example, in the inquiry destination management table 800, the business type (inquiry destination identification character string) “parking management” is associated with the person in charge (contact) “Suzuki” of the business type “parking management business”. registered with Further, in the inquiry destination management table 800, the work type "translation management" and the person in charge of the work type "translation management work" "Tanaka" are registered in association with each other.

As will be described later, by using the inquiry destination management table 800 and the extended answer character string M114 in the answer management table 303, the inquiry destination regarding the assumed question Q51 is specified.

Again, refer to FIG.

In the middle response information A112 (A100) of the response management table 303, in addition to the two extended response character strings M121 and M122, an extended response character string M123 (M100) is further defined. In the extended response character string M123, an inquiry contact identification character string 602 (specifically, the type of work "translation management") for identifying an inquiry contact regarding the assumed question Q52 and attribute information 550 ("Info") are defined in association with each other.

Such an answer management table 303 is registered in the chatbot server 90 . Then, as will be described later, when a single reply message is not extracted from among the plurality of reply messages 400, the computer 50 generates an inquiry destination notification message 700 using the inquiry destination identification character string 600 and sends it to the user. be notified.

FIG. 15 is a flow chart showing the operation of the computer 50 according to the third embodiment. As shown in FIG. 15, in the third embodiment, steps S21 and S22 are executed in addition to steps S11 to S17.

Here, it is assumed that user U100 inputs question message Q11 ("I would like to commute by car") on chat screen 200 (201) (see FIG. 5). However, here, the user U100 (questioner) is an employee seconded from an affiliated company, etc., and none of the three employee-type groups G1 to G3 are assigned to the user U100.

Upon receiving the question message Q11 (step S11), the computer 50 transmits the question message Q11 to the chatbot server 90 (step S12). When the question message Q11 is received, the chatbot server 90 extracts the answer information A111 (FIG. 16) associated with the assumed question Q51 corresponding to the question message Q11 from the answer management table 303 (step S53 (FIG. 7)) Send to computer 50 (step S54).

Then, when the computer 50 receives the answer information A111 (step S13), it acquires question act information (more specifically, employee type information) (step S15). However, here, question act information (classification information) indicating that the user U100 (questioner) does not belong to any of the three employee type groups G1 to G3 (is unclassified) is acquired.

As a result, in the answer message extracting process of step S16, the answer message 400 corresponding to the question act information is not extracted from the plurality of answer messages 411-413 in the answer information A111 (FIG. 16).

Specifically, the computer 50 extracts a plurality of attribute information 500 (501 to 503, 550) using the distinguishing information D10, D20, and from among the extracted plurality of attribute information 501 to 503, 550, the question act information Attribute information 500 corresponding to is extracted. However, here, the question act information indicating that the user U100 does not belong to any of the three employee type groups G1 to G3 (unclassified) is acquired, and the attribute information 500 corresponding to the question act information is extracted. not.

In this case, the computer 50 determines in step S21 that a single reply message 400 is not extracted, and advances the process from step S21 to step S22.

Then, the computer 50 notifies the user U100 (inquirer) of the inquiry destination notification message 700 (see FIG. 17) (step S22). In other words, the computer 50 executes notification processing of the contact notification message 700 as exception processing to the processing (principle processing) of notifying the user of the reply message 400 defined in the reply information A111 in the chatbot server 90. .

Specifically, the computer 50 prepares (stores) a model of the inquiry destination notification message 700 (see FIG. 17) in advance. Computer 50 notifies user U100 of inquiry destination notification message 700 based on the template.

The model of the person-in-charge notification message 700 includes a modifiable part 725 (see FIG. 17) indicating an inquiry destination. In other words, the contact notification message 700 is stored in the computer 50 with the modifiable portion 725 of the contact notification message 700 not entered.

If it is determined in step S21 that a single reply message 400 is not extracted, the computer 50 changes the inquiry contact (contents indicating the inquiry contact) specified using the inquiry contact management table 800 (FIG. 17). By inputting to the part 725, the person in charge notification message 700 is generated, and the inquiry destination notification message 700 is notified to the user (step S22).

Specifically, first, the computer 50 generates an inquiry destination identifying character string 601 (exception processing information) (“parking management”) associated with the attribute information 550 (“Info”) in the response information A111 (FIG. 16). is extracted from the answer information A111.

Then, the computer 50 selects the inquiry destination (“Suzuki” in this case) corresponding to the inquiry destination specifying character string 601 (“parking lot management”) as one of the plurality of inquiry destinations in the inquiry destination management table 800 (FIG. 17). Identify (extract) from within.

In this way, the computer 50 uses the inquiry management table 800 and the extended response character string M114 (attribute information 550, inquiry identification character string 601) in the response management table 303 to solve the assumed question Q51. Identify contact information for inquiries.

After that, the computer 50 changes the changeable part 725 (see FIG. 17) in the template of the person in charge notification message 700 to the contact (contents indicating the contact) "Suzuki", and changes the person in charge notification message 700 (assumed A person-in-charge notification message in response to question Q51) is generated. In other words, computer 50 completes contact notification message 700 .

Then, the computer 50 notifies the user U100 of the generated person-in-charge notification message 700 . Specifically, the computer 50 displays the person-in-charge notification message 700 ("The person in charge is Mr. Prompt U100.

As described above, in the third embodiment, the computer 50 stores a model of the person-in-charge notification message 700, and when a single response message 400 is not extracted from the response information A100, the person-in-charge notification message is generated based on the template. 700 is notified to the user (step S22 (FIG. 15)) (see also FIG. 18). Therefore, even if the answer message 400 corresponding to the question act information is not extracted, the user is notified of a message different from the plurality of answer messages 400 defined in the answer information A100 in the chatbot server 90. , it is possible to respond more flexibly to the question message Q10 from the user.

In this case, the frequency of changes in business personnel (inquiries) within a company is often higher than the frequency of changes in work rules and the like. That is, the change frequency of the person-in-charge notification message 700, which notifies the person in charge of work, is higher than the change frequency of the messages (response messages 411 to 413, etc.) related to work rules and the like.

If such a person-in-charge notification message 700 is registered in the reply management table 300 in the chatbot server 90, the administrator's access frequency to the reply management table 300 is relatively high.

Some conversation processing systems charge for access (communication) to the answer management table 300 . In such a case, it is preferable to restrict access to the reply management table 300 in order to suppress the billing amount.

On the other hand, in the third embodiment, the model of the person-in-charge notification message 700 exists on the computer 50 (conversation processing application) side, and the person-in-charge notification message 700 is generated by the computer 50 and sent to the user. be notified. Therefore, when changing the person in charge, the administrator can enter the name of the person in charge in the inquiry management table 800 (FIG. 17) of the conversation processing application without accessing the response management table 300 in the chatbot server 90. Just change it. Therefore, when the access to the answer management table 300 is charged, the chances of accessing the answer management table 300 can be reduced.

In addition, since the contact management table 800 (FIG. 17) in which personal names (names of persons in charge) are described is stored not in the external chatbot server 90 but in the in-house conversation processing application, personal information leakage is prevented. It is also possible to reduce security risks such as

In the above-described third embodiment, the inquiry destination identifying character string 600 and the attribute information 550 are defined in the response information A100 (A111, A112) (see FIG. 16), but the inquiry destination is not limited to this. Neither the specifying character string 600 nor the attribute information 550 need be defined in the answer information A100. In this case, when a single reply message is not extracted from the plurality of reply messages 400, a message (also called a preliminary reply message) provided in advance in the computer 50 is notified to the user (questioner). may Specifically, a message common to each of the assumed questions Q51 and Q52 (such as "Please ask the person in charge.") may be notified to the user (questioner).

Further, here, the third embodiment is exemplified as a modification of the first embodiment, but it is not limited to this, and the idea of the third embodiment may be applied to the second embodiment. .

For example, in the second embodiment, if the question message Q10 (Q14 (see FIG. 12)) regarding the date of the next power outage is accepted after August 18, 2018, the answer message 400 corresponding to the question date information is the answer. Not present in information A104 (see FIG. 13), no single reply message is extracted. In such a case, the computer 50 may generate a person-in-charge notification message 700 and notify the user (questioner).

In addition, in the third embodiment, the contact management table 800 (FIG. 17) is stored in the computer 50. However, the contact management table 800 is not limited to this, and the contact management table 800 can be stored in a device other than the computer 50 (chatbot). server 90, etc.).

<4. Modifications, etc.>
Although the embodiments of the present invention have been described above, the present invention is not limited to the contents described above.

For example, in each of the above-described embodiments and the like, the reply message 400 and the attribute information 500 are defined separately (separately) in the reply extended character string M100 (see FIG. 8, etc.), but the present invention is not limited to this. , the reply message 400 and the attribute information 500 may be defined without distinction in the reply extended character string M100.

FIG. 19 shows the reply management table 300 (304) according to this modified example.

In the answer management table 304, the attribute information 500 is included in the answer message 400 in each of the answer information A121 and A122 corresponding to each assumed question Q51 and Q52, and the answer message 400 and the attribute information 500 are not distinguished. , an extended type character string for answer M100 is configured.

For example, in the extended reply character string M171 (M100) in the reply information A121, the attribute information 571 (character string "full-time employee who has been with the company for two years or more") indicating the employee type group G1 is replaced by the reply message 451 (employee type group message for users belonging to G1). In addition, in the extended response character string M172 (M100) in the response information A121, the attribute information 572 (character string "full-time employee in the first year of employment") indicating the employee type group G2 is replaced by the response message 452 (employee type group G2 messages for users belonging to ). Further, in the extended response character string M173 (M100), the attribute information 573 (character string "temporary employee") indicating the employee type group G3 is included in the response message 453 (message for users belonging to the employee type group G3). is

When the question message Q11 (see FIG. 5) from the user is accepted (step S11), the chatbot server 90 sends the answer information A121 (FIG. 19) associated with the assumed question Q51 corresponding to the question message Q11 to the computer. 50 (step S54 (FIG. 7)).

When the computer 50 receives the answer information A121 from the chatbot server 90 (step S13 (FIG. 6)), it acquires question act information (more specifically, employee type information) (step S15).

FIG. 20 shows a classification information management table 900 (902) according to this modified example. In the classification information management table 902, user information of each user and a character string indicating the employee type of each user are associated and registered. Based on the user information (userU1) of the questioner (for example, user U1), the computer 50 acquires the questioner's employee type information (character string "first year full-time employee") as question act information (step S15 ).

Then, the computer 50 executes reply message extraction processing based on the employee type information and the reply information A121 (see FIG. 18) from the chatbot server 90 (step S16). Specifically, the computer 50 selects an answer including the employee type information (character string “first year full-time employee”) from among the multiple answer extended character strings M171 to M173 included in the answer information A121. The expanded character string for answer (here, the expanded character string for answer M172) is extracted (step S16). Here, a line feed code (not shown) is embedded as distinguishing information D10 at the end of each extended answer character string M170. Assume that the character strings M171 to M173 are extracted from the answer information A121 while being distinguished from each other.

Then, the computer 50 notifies the user U1 (questioner) of the extracted expanded answer character string M172 (that is, the answer message 452) (step S17).

In this manner, the reply message 400 and the attribute information 500 may be defined without distinction in the reply extended character string M100.

Further, in each of the above-described embodiments and the like, conversation processing in a text chat format in which conversation (chat) is performed with the user using text messages is exemplified, but the present invention is not limited to this. Conversation processing may be performed in a voice chat format in which a conversation (chat) is performed with.

1 conversation processing system 50 computer (information processing device)
90 chatbot server 200 chat screen 300 answer management table 400 answer message 500 attribute information A100 answer information Q10 question message

Claims (17)

A conversation processing system,
an information processing device;
a server having a database in which assumed questions, which are assumed questions, and answer information corresponding to the assumed questions are associated and defined;
with
The information processing device is
receiving means for receiving a question message from a user;
sending means for sending the question message to the server;
has
When the server receives the question message from the information processing device, the server extracts one answer information associated with one assumed question corresponding to the question message from the database and transmits the answer information to the information processing device. ,
The one answer information defines a plurality of answer messages for the one assumed question,
The information processing device is
receiving means for receiving the one reply information from the server;
Acquisition means for acquiring question act information, which is information relating to the question act by the user;
Extraction means for extracting a single answer message corresponding to the question act information from among the plurality of answer messages in the one answer information;
notification means for notifying the user of the single reply message;
further having
each of the plurality of reply messages is associated with at least one piece of attribute information indicating mutually different attributes of each reply message in the one reply information;
The extraction means extracts the attribute information corresponding to the question act information from the one answer information, and extracts the single answer message associated with the attribute information corresponding to the question act information from the plurality of answers. A conversation processing system characterized by extracting from messages. The conversation processing system of claim 1 , wherein
The information processing device stores a template of an inquiry destination notification message, which is a message for notifying an inquiry destination regarding the one assumed question,
The conversation processing system, wherein the notifying means notifies the user of the contact notification message based on the template in the information processing device when the single reply message is not extracted by the extracting means. . A conversation processing system according to claim 2 ,
In the information processing device, a plurality of inquiry destinations and a plurality of identification character strings for identifying each of the plurality of inquiry destinations are registered in association with each other,
the template of the inquiry destination notification message includes a predetermined changeable part indicating the inquiry destination;
The notification means, if the single reply message is not extracted by the extraction means,
Extracting one specifying character string, which is one of the plurality of specifying character strings, from the one answer information, and one inquiry contact corresponding to the one specifying character string from among the plurality of inquiries,
A conversation processing system, wherein said predetermined changeable portion of said template is changed to said one inquiry destination, said inquiry destination notification message is generated, and said inquiry destination notification message is notified to said user. In the conversation processing system according to any one of claims 1 to 3 ,
A conversation processing system, wherein the question act information includes classification information given to a questioner who has performed the question act. In the conversation processing system of claim 4 ,
The conversation processing system, wherein the classified information is information classified based on at least one of employment status and length of service at the company to which the questioner belongs. In the conversation processing system according to any one of claims 1 to 3 ,
A conversation processing system, wherein the question act information includes question date information indicating a date on which the question act was performed. In the conversation processing system according to any one of claims 1 to 3 ,
A conversation processing system, wherein the question act information includes question time information indicating a time when the question act is performed. In the conversation processing system according to any one of claims 1 to 3 ,
A conversation processing system, wherein the question act information includes question place information indicating a place where the question act is performed. In the computer built into the information processing equipment,
a) accepting a question message from a user;
b) a step of transmitting the question message to a server having a database that defines an assumed question that is an assumed question and answer information corresponding to the assumed question in association with each other;
c) a step of receiving, from the server, one answer information associated with one assumed question corresponding to the question message and defining a plurality of answer messages for the one assumed question; When,
d) obtaining question act information, which is information about the question act by the user;
e) extracting a single answer message corresponding to the question act information from among the plurality of answer messages in the one answer information;
f) notifying said user of said single reply message;
A program that executes
each of the plurality of reply messages is associated with at least one piece of attribute information indicating mutually different attributes of each reply message in the one reply information;
In step e), the attribute information corresponding to the question act information is extracted from the one answer information, and the single answer message associated with the attribute information corresponding to the question act information is extracted from the plurality of answer information. A program characterized in that it is extracted from the answer message of In the program according to claim 9 ,
The information processing device stores a template of an inquiry destination notification message, which is a message for notifying an inquiry destination regarding the one assumed question,
In the step f), if the single reply message is not extracted in the step e), the user is notified of the contact notification message based on the template in the information processing device. program to do. In the program according to claim 10 ,
In the information processing device, a plurality of inquiry destinations and a plurality of identification character strings for identifying each of the plurality of inquiry destinations are registered in association with each other,
the template of the inquiry destination notification message includes a predetermined changeable part indicating the inquiry destination;
In said step f), if said single reply message was not extracted in said step e):
One specifying character string, which is one of the plurality of specifying character strings, is extracted from the one answer information, and one query corresponding to the one specifying character string A destination is identified from among the plurality of inquiries,
A program, wherein the modifiable portion of the template is changed to the one inquiry destination, the inquiry destination notification message is generated, and the inquiry destination notification message is notified to the user. In the program according to any one of claims 9 to 11 ,
A program, wherein the question act information includes classification information given to the questioner who performed the question act. The program according to claim 12 ,
A program, wherein the classified information is information classified based on at least one of employment status and length of service at the company to which the questioner belongs. In the program according to any one of claims 9 to 11 ,
A program, wherein the question act information includes question date information indicating a date on which the question act was performed. In the program according to any one of claims 9 to 11 ,
A program, wherein the question act information includes question time information indicating a time when the question act is performed. In the program according to any one of claims 9 to 11 ,
A program, wherein the question act information includes question place information indicating a place where the question act is performed. An information processing device,
receiving means for receiving a question message from a user;
transmission means for transmitting the question message to a server having a database in which an assumed question, which is an assumed question, and answer information corresponding to the assumed question are defined by associating them;
receiving means for receiving, from the server, one piece of answer information associated with one assumed question corresponding to the question message and defining a plurality of answer messages for the one assumed question; ,
Acquisition means for acquiring question act information, which is information relating to the question act by the user;
Extraction means for extracting a single answer message corresponding to the question act information from among the plurality of answer messages in the one answer information;
notification means for notifying the user of the single reply message;
with
each of the plurality of reply messages is associated with at least one piece of attribute information indicating mutually different attributes of each reply message in the one reply information;
The extraction means extracts the attribute information corresponding to the question act information from the one answer information, and extracts the single answer message associated with the attribute information corresponding to the question act information from the plurality of answers. An information processing device characterized by extracting from a message.

Images